Optimising time varying reserve strategies for conversation biology

How can we better protect our marine ecosystems in a dynamic and complex environment?

Summary

The design of marine reserves is a complex and important area of conservation biology, balancing ecological objectives against the effects of restricted use on other activities such as fishing and tourism. Data science can offer detailed insights on these kinds of issues. For example, the most recent Great Barrier Reef rezoning used the technique of simulated annealing to determine an optimum location profile.
However, extant tools only answer a static question: “what reserve locations maximise my objectives, right now?” We can apply these techniques to projected future climates, but no tools exist which properly deal with the dynamic and uncertain nature of this problem.
Our project aims to rectify this. Using the techniques of binary convex quadratic programming, we will develop tools to design time-varying reserves, incorporating spatially varying profiles of risk and reward, and the many competing objectives at play. As the effects of climate change become increasingly pronounced, this problem is only becoming more acute. Governments need to plan now for the uncertain futures of our marine habitats, and we hope to contribute to this effort.

Aim & Methodologies

The initial stage of this project aims to explore the boundaries of this problem, and to test the capabilities of current optimisation software. Supported by the seed funding grant, we will:

  • Compute several optimum reserve networks in a case study, over the years 2030-2100, accounting for change and uncertainty due to climate change. This will allow us to assess how the needs of our marine ecosystems might be expected to change in the coming decades.
  • Develop a multi-period extension of a current model, which has been used by RR in recent research. In doing so we will determine the capabilities and limitations of general-purpose solvers applied to this problem.
  • Produce visualisations to communicate potential time-varying reserve schemes.

We hope that this project will allow us to develop a grant application for a fully-fleshed-out project, and that it will initiate a new and productive collaboration between the School of Mathematics and Statistics, the Melbourne Centre for Data Science and the School of Geography, Earth and Atmospheric Sciences.

Researchers

Chief Investigator
  • Dr Chris Baker, Research Fellow - Statistics For Biosecurity Risk
    School of Mathematics and Statistics
    Faculty of Science
  • Co-Investigators
    • Dr Rebecca Runting, Arc Decra Fellow
      School of Geography, Earth and Atmospheric Sciences
      Faculty of Science
    • Dr Matthew Tam, Decra Senior Fellow
      School of Mathematics and Statistics
      Faculty of Science
    • Tom Waring, Research Assistant
      School of Mathematics and Statistics
      Faculty of Science